Matia Lineage Now Runs Through Your Snowflake Transformations

Matia now traces lineage through transformations you run inside Snowflake, including stored procedures and dynamic tables. Through Matia MCP, Cortex Code (CoCo) and Cortex Agents can use it too.

It's rolling out now for Snowflake Enterprise edition and above.

Snowflake transformations in Matia lineage

A lot of data teams transform inside Snowflake. A stored procedure merges yesterday's orders into a reporting table. A dynamic table keeps a daily revenue rollup fresh. Until now, those steps were a gap in Matia's lineage graph. You could see the table a procedure wrote to, but not how the data got there.

Matia now reads lineage directly from Snowflake Horizon, with no SQL parsing on our side. Each transformation shows up as its own node between the tables it reads and writes. Column-level edges run through it, so you can follow a single field from source to output. Views and dynamic tables get their own badges, which makes a transformation output easy to tell apart from a plain table.

Impact analysis runs through these nodes too. Change a source table and Matia shows what sits downstream, including everything the procedure feeds.

Check impact before CoCo changes a transformation

MCP is an open standard, so Matia MCP works in any client that supports it. CoCo and Cortex Agents both do.

Say an engineer asks CoCo to rewrite the procedure that builds fct_orders. Before it touches the code, CoCo can ask Matia what depends on that table. Matia comes back with the monitors watching it and the reverse ETL sync that pushes it to Salesforce. It also lists the dashboards that read from it. The engineer approves the change knowing exactly what it touches.

Matia MCP recommends. You confirm every action before anything happens.

Trace a bad answer back to its source

Now the other direction. A Cortex Agent answers a revenue question with a number that looks off. The table it queried was built by a stored procedure.

Matia follows the path back from the output table, through the procedure, to the source table and the Matia sync that loaded it. That's where it finds the freshness monitor that failed at 8 a.m. The agent gets root cause and downstream impact in one answer, instead of piecing it together from separate tools. In our testing, workflows that took an agent ten or more tool calls against single-vendor MCPs took one or two against Matia.

When a question needs live numbers, Matia calls Snowflake MCP and combines the results with its own context.

The operational layer around Snowflake

Matia is the operational data infrastructure around Snowflake. It brings ETL, Reverse ETL, Observability, and Catalog together in one Unified DataOps Platform. Data comes in reliably and gets checked before it lands. Then it goes back out to the tools your business runs on. Snowflake stays your source of truth.

Ramp runs its Postgres-to-Snowflake replication through Matia. Historical resyncs that used to take 3.5 days now take 4 hours.

Matia is available on Snowflake Marketplace. Find us there, or book a demo to see your Snowflake transformations in Matia lineage.